Re: Experienced Statistician to help decide whether a regression is legitim
- From: "Reef Fish" <large_nassua_grouper@xxxxxxxxx>
- Date: 27 Oct 2006 17:33:03 -0700
Anahita wrote:
I didn't realize how long a single conversation we had! For this
reply, I'll cut out
MOST of what we had discussed already and comment only on a few.
.. however as a devils advocate... some malpractice is really based on "ignorance" (which is bliss!)- and this ignorance is "good faith"...
Most malpractice ARE based on "ignorance", but I was talking about
my experience in THIS group, that after the malpractice had been
explained and re-explained a dozen times, the same person
continues to argue that it's not malpractice -- that's ignorance of
a different kind -- refusal to LEARN, because I always point to
reputable books and articles that corrects the malpractice, but
the few frequent malpractioners in sci.stat.math seldom ever
read those references and continued to wallow in their BLISS of
ignorance.
C'est la difference. That's what you say in France. ;)
Yes then one should dig and search.. but not everyone has the mindset of a "researcher"...
Some of this malpractice is a consequence of a lack of modesty as well...
Most of the malpractice I found here are NOT in research, but
merely executing simple applied tasks that can hardly be
considered research. So it really are deficiencies that resulted
in malpractice. I think what I've seen here (for the most part)
are beyond a lack of modesty.
There have been Executive Programs in Business
Schools throughout the country, in their MBA programs
in which ONLY high level managers and
company CEOs are admitted. They are taught not only
how to recognize what is good and what is bad, but
actually DO many of the analysis themselves.
I had no knowledge of that.. and if this is true this is wonderful..
I was aware of this kind of program from my colleagues at the
University of Chicago, Harry Roberts in particular, who told me
many of the things the Executives do in their classes. I didn't
teach any of the Exec Program courses until 1982 when I was
a Visiting Prof at the MBA Exec Program at Vanderbilt. It was
a most rewarding experience to find the eager-to-learn EXECS
(most CEO and high level managers) who were NOT the
brightest students in statistics, yet they were genuinely interested
in learning, and they show their appreciation and above all, they
don't blame what they don't grasp on ME (as most of the
undergrads do). That was the only time in my academic career
that I had an Exec student writing me after the course was all
over and he received a D how he appreciated my efforts. In
some way, that was A high point of my career, tobe genuinely
thanked by a student who made a "D" in my course. Elsewhere
I've been taken to the bonfire for lynching when the students
did not get the grade they thought they deserve but didn't.
It makes a whole world of difference between the RESPONSIBLE
execs (which is why they got to their positions in the first place)
and the irresponsible brats (now common in all colleges). That
group of MBA Execs invited me back to their Reunion (at their
expense, and I was the only one so invited out side of Vanderbit)
5 years later, when I was a Visiting Prof at Harvard.
however I see a few points I can state here:
first it does take quite a lot of perspiration to become a mathematician or a statistician.. so going through a few
courses if it might enlighten you somewhat doesn't make you all of a sudden a "magus of statistics"..
Of course not. But at the hands of good profesors such as Harry
Roberts and a few others, they are orders of magnitude better than
those managers who had NOT been through the few courses of
education on how statistics is properly used.
Second those managers who have/take their spare time to go through these courses might be mostly those who
think it matters... (selection bias?)...
Possibly. But if they are already CEOs, it becomes mostly one of
the desire to learn and be a BETTER CEO. There are always
exceptions of course.
So even there.... there would be matter for debate or?
My >co-author and former colleague
Harry V. Roberts of
the University of Chicago had been a strong moving
force in that
program
at Chicago as well as in the movement to educate the
Management.
Send him my belated congratulations for his dedication.
Perhaps you didn't have a web brower to read what's in the URL.
I regret to say that it is too belated. Because the page was an
obiturary of his death in 2004. That was during the YEARS of
my complete disassociation with ALL statisticians and ALL statistics
(from about 2000 to 2005 when I re-started in sci.stat.math) because
of my disgust with what the Educational Institutions did to reward
the professors who did not TEACH but gave out high grades --
and they are the CAUSE of so many mal-practice statisticians we
see today. The statistics professor did LITTLE if ANY to try to
right the wrong.
So, it was my deepest regret that I didn't even know John Tukey
died until two years later, and that I didn't know Harry died until
fairly recently.
The whole thing is certainly a good idea and shows he was persuasive enough.
He was a remarkable human being. The webpage highlighted
that he was a Bayesian which not many people knew. I've
stated several times in this group that I learned my Bayesian
Statistics from L.J. Savage, Harry Roberts, and few other
very reputable Bayesians when some Illywhacker and other
complete NON-Bayesians made all kinds of noise challenging
me about my knowledge about Bayesian stat. when they were
the ones who were completely bankrupt in said knowledge.
http://chronicle.uchicago.edu/041007/obit-roberts.shtml
He
even taught principles management that are much more
management than
statistics.
That is the page about Harry Roberts. Perhaps that line
was truncated so that you couldn't access it because I
can see the "l"of shtml dangling below. :0)
l
#> Toward the end of his career, Roberts helped
develop a curriculum in
#> Total Quality Management at the GSB.
Harry even had succeess in pursuading several major
companies to REQUIRE their upper level management positions to be
filled by only those who had several advanced courses in statistics.
Here, with all du e respect, are you sure that this is a good criterion? After all division of labor is about everybody should be doing that for which he is best qualified.. I guess managers should manage and statistician "statisticize"
Yes, but in an industry which so much depends on the research and
study BASED on statistical data, such a requirement would not TAX
any high level manager -- management is a skill that can almost be
learned WITHOUT any formal study whereas statistics is a skill
that CANNOT be learned without a very carefully supervised LEARNING
experience.
hmmm that was my reason to leave academia... this "tolerant" attitude to "low or nonsensical" research ...
Me too. So, we have at least that much IN COMMON for out
reasons to leave academia. :-) Publish or perish is ALMOST
as bad as "Give grades to please the students because they
pay the tuition -- forget about teaching them anything."
They are much more ready to
criticize ME for pointing out the malpractice than
criticizing those who malpractice.
I have not followed the debate. It is thus a courageous attitude of yours to defend and try to point out malpractice... one is seldom rewarded for that. It takes on top a communicator talent
That is not restricted to any particular debate. That has been
the PREVAILING atmosphere ever since I stepped in sci.stat.math
about 5,000 posts ago. :-)
But being human (or pescan?)
Ah, that's Spanish for "the fish", and I thought you were French.
For a second, I thought you were referring to some poster
who pasted with the name "pescaand" who first thanked me
profusely in 2005 for being so patient in teaching her/him
about what a Linear Model was, and then suddenly re-appeared
in 2006 to post the most bitter, unwarranted, and vicious
attack on me because I had made a passing remark about a
UK student who was apparently his/her girlfriend/boyfriend,
who had dropped out of that discussion long ago.
you know very well how difficult to us all it is to admit that one errs....
It is VERY easy for me because I seldom err, and if I do, and was
pointed out, I immediately accept it. If you can't do that, then you
lose ALL credibility.
Even some of my CRITICS (of my no-nonsense way of blunt
communication as is my trademark in academia) readily point
out to others my trademark phrase "I stand corrected" on those
FEW occasions I did err and was pointed out by others, whether
the error was large or small.
so then Mr. Reef Fish, humans remain humans even if dealing with statistics..you cannot expect that those you enlighten do not go through the tortures of realizing that what they believed they knew might have benn illusory..
It becomes a torture ONLY when they refuse to LISTEN and at
least TRY to understand WHERE they erred. In m00es's case,
which was an "extremely" of such a case, he couldn't wait when I
said "read it in my next post" (and he had already made the SAME
NOISE posting exactly the same that he had erred).
Those who continued arguing when they SHOULD have known
how they erred are the ones who brought torture to themselves.
In the end, instead of losing ONCE (admit and move on), they
lose 100 TIMES and leave a permanent record in the archives
of their ignorance and obnoxiousness, not to mention the
pollution of an entire newsgroup.
It is more difficult to be a top level APPLIED
statistician than a brain surgeon. If a brain surgeon is
obviously incompetent, his patients DIE and is easily observable. :-)
ahahah....
Thank you for being a receptive audience on that. :)
When applied statisticians
malpractice and have strong influence on CEOs of
companies, the effect is not as readily visible.
Companies ALWAYS have the option to hire competent
statisticians to help do the job their own in-house statistics staff
is incapble of handling properly or well.
Well there you are... they do in theory, but in practice.. it is slightly more complex... first you have to understand what kind of specialist you hire, and then screening humans is difficult you know that as well..
That's why the EXEC program CEOs and upper level managers are
in a better position to know whom to hire.
and then it takes time to discover that your team doesn't know how to work ( especially if you have not been through the courses listed above...).. which is the case in the issue I am talking about... and then if the hired consultants also fail to "see" the problem...
All of that is certainly possible! The more ignorant the management,
the more probable. :-)
C'est la vie. We are not living in Utopia. We have both seen the
Real World. Kicking a strawman is easy to do. Recognizing and
appreciating a solid statistician is a much harder task.
Another French say says: "avec des si on met Paris en bouteille" meaning that whith a causal chain of "if2s you can even bottle "Paris".. i.e. manage the impossible..
That sounds like my mentor's favorite saying against Fisher's
idea of randomization: "What if you throw up a group of blocks
of letters and it fell spelling 'C H R I S T M A S'"?
One of the things one learns is how to recognize a
BAD book.
agreed....
In any case thank you for your refreshing thoughts and the story of the courses for manager.. the US is always in advance!
Touche. The US is in advance in some small areas. The US
remains foreever behind in the math and science areas in spite
of having the best in the same areas. That's because the best
FEW are at the tail of the dog. They cannot wag the dog, the
ever content we-are-the-best majority even if they ranked FAR
behind most countries, even a few in the 3rd world, on math
related areas.
-- Reef Fish Bob.
.
- Follow-Ups:
- References:
- Prev by Date: Re: Testing for normality
- Next by Date: Re: Experienced Statistician to help decide whether a regression is legitim
- Previous by thread: Re: Experienced Statistician to help decide whether a regression is legitim
- Next by thread: Re: Experienced Statistician to help decide whether a regression is legitim
- Index(es):
Relevant Pages
|
